29-05-2012, 12:43 PM
SUITABILITY STUDY OF DS-UWB AND UWB-FM FOR MEDICAL APPLICATIONS
[attachment=23317]
ABSTRACT
In this paper, the suitabilities of direct sequence ultra
wideband (DS-UWB) and ultra wideband frequency
modulation (UWB-FM) are studied for medical applications.
The system performances are studied by applying the
measured hospital channels. The channel measurements were
done at the Oulu University hospital in 2005. The
measurement campaign included three different hospital
environments, i.e. the operating room, intensive care unit and
x-ray operating room. The measured frequency range was
from 3.1 GHz to 6.0 GHz. The simulation results indicate that
UWB-FM outperforms DS-UWB with low data rates, i.e. less
than 1 Mbps, due to the simpler implementation. When data
rate increases, DS-UWB is reasonable choice for medical
applications.
I. INTRODUCTION
Nowadays, cables attaching patient monitoring sensors to
monitoring devices disturb and obstruct nursing staff. Vital
parameters, such as blood pressure, electrocardiography
(ECG), respiration rate, heart rate and temperature, from
patient who is in critical condition are measured all the time.
Wired connections make free redeployment of patient
between separate units more difficult. In addition, wires
complicate the movement of the patient. Wireless connections
will offer quick and easy way to redeploy the patient, e.g.,
from an operating room to an intensive care unit (ICU) and
make the movement of patient more easier.
The current wireless techniques applied in hospitals are
based on wireless medical telemetry service (WMTS) and
wireless local area network (WLAN) standards [1].
Nevertheless, data transmission from patient monitoring
sensors to monitoring devices has no viable solution. The
potential solution for short range communication in the
hospitals can be found from the wireless body area network
(WPAN) standards, e.g., ultra wideband (UWB), ZigBee and
Bluetooth [1].
In this paper, two singleband UWB systems, i.e., direct
sequence UWB (DS-UWB) and UWB frequency modulation
(UWB-FM), are studied in the measured hospital channels. In
the traditional DS-UWB technique, the energy of the
information signal is spread with a pseudo random code
sequence. When a very narrow chip waveform is applied, it
inherently generates ultra wide spectrum [2]. The UWB-FM
technique is based on the dual frequency modulation [3].
These systems are discussed in Section II in more details.
The channel measurements from 3.1 to 6.0 GHz were
carried out at the Oulu University hospital [4]. At the time of
the measurements, the Federal Communications Commission
(FCC) was the only governing body that allowed UWB
system to operate in the frequency range from 3.1 to 10.6
GHz [5]. Therefore, the studies are focused on the UWB
frequency range defined by the FCC. The measurements were
carried out in three environments; an operating room, an
intensive care unit and an x-ray examination room.
The paper includes the following paragraphs; Section II
presents the system models and channel models. In Section
III, the simulation parameters and configurations are
discussed. The results are presented in Section IV. In the end,
the paper is concluded by Section V.
II. SYSTEM MODELS
In this section, the system models for DS-UWB, UWB-FM
and hospital channel models used at the simulations are
discussed.
A. DS-UWB
In the DS-UWB technique, the energy of the information
signal is spread by utilizing pseudo random codes. When the
very narrow pulse is applied as a chip waveform, the energy
of the signal is spread over ultra wide spectrum. At a receiver,
a selective rake receiver is applied. The rake receiver
improves the performance by utilizing time diversity, i.e. by
capturing the signal energy propagated through a channel by
different paths and combining the signal components. [2]
Maximum ratio combining (MRC) is an optimal combining
technique by weighting received signal components with
estimated received amplitude and then coherently combining
them. The decision variables in MRC is given by [2]
( τ ) ( ) , 0,1,
1 0
*
,MRC
b
=Σ ∫ − =
=
U a r t w t dt i n i
N
n
T
i n
(1)
where n is the multipath component, an is the complex gain of
the nth multipath component, Tb is the duration of the data bit,
r(t) is the received signal, t is time, τn is the delay of the nth
multipath component and wi(t) is the pulse waveform of the
data bit i.
Equal gain combining (EGC) is similar to MRC, but it does
not weight the signal components with the estimated received
amplitude [2]. Square-law combining (SLC) is non-coherent
approach, where the phase of the signal is considered to be
unknown. This leads to simpler implementation of the
receiver with the cost of the performance. The decision
variables in SLC are [2]
( τ ) ( ) , 0,1.
1
2
0
,SLC
b
=Σ ∫ − =
=
U r t w t dt i
N
n
T
i n i
(2)
The 11th International Symposium on Wireless Personal Multimedia Communications (WPMC 2008)
B. UWB-FM
UWB-FM is a simple technique for UWB communication
with low data rates. It applies a double frequency modulation
(FM): digital frequency shift keying (FSK) with a low
modulation index followed by analog FM with a high
modulation index. This yields to the FM signal V(t) [3]:
( ) sin(ω βcos(ω ) φ ), c m 0 V t = A t − t + (3)
where A is the amplitude, ωc is the angle frequency of the
carrier signal, β is the modulation index, ωm is the angle
frequency of the modulating signal and φ0 is the arbitrary but
time-independent constant.
At the receiver, the FM signal is demodulated with a delayline
FM demodulator. This approach does not need any
frequency translation. Therefore, no local oscillator and no
carrier synchronization are required. This reduces the
complexity of the transceiver. In the delay-line FM
demodulator, the received signal is delayed with time that is
equal to an odd multiple of a quarter period for the carrier
frequency of the FM signal. FSK demodulation can be done
with a bandpass filter followed by a phase-locked loop. [3]
C. Hospital Channel Models
The channel measurement campaign was carried out at the
Oulu University hospital in 2005 [4]. The measurements were
done in three different environments; an operating room
(OR), an x-ray examination room (XR) and an intensive care
unit (ICU) in the frequency range from 3.1 GHz to 6.0 GHz.
The environments were static during the measurements, i.e.
there were no movement inside a room when recording the
data, expect in the ICU. The multipath model was obtained by
investigating the multipath propagated signals in the power
delay profile (PDP). The model was shown to be similar as a
modified IEEE 802.15.3a channel model defined in [6]. The
modified IEEE 802.15.3a parameters for the hospital
environments are presented in Table 1. The example channel
realizations for the environments are illustrated in Figure 1.
The parameters occurred in Table 1 are [6]
• Λ = Cluster arrival rate,
• λ = Ray arrival rate,
• Γ = Cluster decay factor,
• γ = Ray decay factor,
• σ1 = Standard deviation of cluster lognormal fading term,
• σ2 = Standard deviation of ray lognormal fading term and
• σx = Standard deviation of lognormal shadowing term for
total multipath realization.
Table 1: Modified IEEE 802.15.3a model parameters for
hospital.
Model
parameters
Operating
room (OR)
X-ray
room (XR)
Intensive care
unit (ICU)
Λ [1/ns] 0.04 0.05 0.09
λ [1/ns] 2 1.5 2
Γ [ns] 9 13.3 16
γr [ns] 8 10 5
σ1, σ2 [dB] 3.4 3.4 3.4
σx [dB] 1.5 1.5 1.5
Figure 1: One channel realization for each environment.
D. Applications
Electrocardiography (ECG) is an essential tool for
investigating cardiac arrythmias. It is also useful in
diagnosing cardiac disorders such as myocardial infarction
[7]. The standard ECG measurement contains information
from 12 leads. The six chest leads (V1 to V6) measure the
activity of the heart in the horizontal plane. The activity in the
vertical plane is measured by the six limb leads (I, II, III,
aVR, aVL and aVF). When the sample rate of 1250 samples
per second and the resolution of 12 bits/sample are applied,
the total data rate (Rb) from the ECG measurement is 180
kbps [8].
The physiological properties of the muscles at rest and in
contraction are evaluated and recorded by using a technique
called electromyography (EMG). In a surface EMG,
measurements are done non-invasive and it can be applied to,
e.g., childbirth to measure the contraction intervals. In the
surface EMG, the signals from reference and two detection
electrodes construct the EMG signal [9]. Each of the electrode
produces the information rate of 600 kbps with the sample
rate of 50 samples per second and the resolution of 12 bits per
sample [8]. Hence, the total data rate is 1.8 Mbps.
A possible application for x-ray examination is to transmit
wirelessly the taken images from a measurement device to a
viewer, e.g., to computer. Since uncompressed x-ray images
are very large, the wireless link should be fast enough. The
The 11th International Symposium on Wireless Personal Multimedia Communications (WPMC 2008)
data rate for the x-ray imaging application is chosen to be 24
Mbps.
III. SIMULATION CONFIGURATIONS
In order to evaluate the suitability of the DS-UWB and UWBFM
systems for the link between the patient monitoring
sensors and the monitoring devices, software simulators were
implemented in MatlabÓ. In this section, the simulation
parameters and configurations are presented and justified. The
parameters for the systems are chosen so that the systems are
occupying approximately the same frequency band. The
spectrum allocations of the systems with defined parameters
are depicted in Figure 2.
Figure 2: Spectrum allocations of UWB-FM and DS-UWB
with defined parameters.
The chip waveform for DS-UWB was chosen to be 10th
derivative of the Gaussian monocycle with the pulse length
(Tp) of 0.7 ns, thus having the center frequency (fc) of 4.57
GHz and bandwidth (W) of 3.11 GHz. The studied receiver
algorithms for DS-UWB are chosen to be coherent MRC and
EGC and non-coherent SLC with 8-finger selective rake
receiver. The selective rake receiver with 8 fingers has been
shown to be adequate in order to trade-off between
complexity and performance [10]. The modulation scheme for
coherent detection algorithms is binary pulse amplitude
modulation (BPAM) and on-off keying (OOK) for noncoherent
detection algorithm, respectively. BPAM cannot be
applied for SLC, because the polarity of the data bit is needed
in the decision. In OOK, nothing is transmitted in the case of
bit "0". The pulse power of OOK is twofold compared to
BPAM to attain same average transmitted power. According
to the data rates, presented in Section II, pulse repetition
processing gains are adjusted. The pulse repetition processing
gain in decibels is defined as
Therefore, the processing gains for 1.8 Mbps and 24 Mbps are
29.0 dB and 17.7 dB, respectively. With the data rate of 180
kbps and the pulse length of 0.7 ns, the processing gain is as
much as 39 dB. This requires a huge calculation capacity, and
therefore it is not feasible for the medical sensor applications.
In the UWB-FM technique, the channel bandwidth is
adjusted with the modulation index of FM. The UWB-FM
system is studied with the channel bandwidths of 500, 1000,
2000 and 2900 MHz. The centre frequency of the carrier is
4.55 GHz. Through the simulations, the modulation index of
FSK is set to one.
[attachment=23317]
ABSTRACT
In this paper, the suitabilities of direct sequence ultra
wideband (DS-UWB) and ultra wideband frequency
modulation (UWB-FM) are studied for medical applications.
The system performances are studied by applying the
measured hospital channels. The channel measurements were
done at the Oulu University hospital in 2005. The
measurement campaign included three different hospital
environments, i.e. the operating room, intensive care unit and
x-ray operating room. The measured frequency range was
from 3.1 GHz to 6.0 GHz. The simulation results indicate that
UWB-FM outperforms DS-UWB with low data rates, i.e. less
than 1 Mbps, due to the simpler implementation. When data
rate increases, DS-UWB is reasonable choice for medical
applications.
I. INTRODUCTION
Nowadays, cables attaching patient monitoring sensors to
monitoring devices disturb and obstruct nursing staff. Vital
parameters, such as blood pressure, electrocardiography
(ECG), respiration rate, heart rate and temperature, from
patient who is in critical condition are measured all the time.
Wired connections make free redeployment of patient
between separate units more difficult. In addition, wires
complicate the movement of the patient. Wireless connections
will offer quick and easy way to redeploy the patient, e.g.,
from an operating room to an intensive care unit (ICU) and
make the movement of patient more easier.
The current wireless techniques applied in hospitals are
based on wireless medical telemetry service (WMTS) and
wireless local area network (WLAN) standards [1].
Nevertheless, data transmission from patient monitoring
sensors to monitoring devices has no viable solution. The
potential solution for short range communication in the
hospitals can be found from the wireless body area network
(WPAN) standards, e.g., ultra wideband (UWB), ZigBee and
Bluetooth [1].
In this paper, two singleband UWB systems, i.e., direct
sequence UWB (DS-UWB) and UWB frequency modulation
(UWB-FM), are studied in the measured hospital channels. In
the traditional DS-UWB technique, the energy of the
information signal is spread with a pseudo random code
sequence. When a very narrow chip waveform is applied, it
inherently generates ultra wide spectrum [2]. The UWB-FM
technique is based on the dual frequency modulation [3].
These systems are discussed in Section II in more details.
The channel measurements from 3.1 to 6.0 GHz were
carried out at the Oulu University hospital [4]. At the time of
the measurements, the Federal Communications Commission
(FCC) was the only governing body that allowed UWB
system to operate in the frequency range from 3.1 to 10.6
GHz [5]. Therefore, the studies are focused on the UWB
frequency range defined by the FCC. The measurements were
carried out in three environments; an operating room, an
intensive care unit and an x-ray examination room.
The paper includes the following paragraphs; Section II
presents the system models and channel models. In Section
III, the simulation parameters and configurations are
discussed. The results are presented in Section IV. In the end,
the paper is concluded by Section V.
II. SYSTEM MODELS
In this section, the system models for DS-UWB, UWB-FM
and hospital channel models used at the simulations are
discussed.
A. DS-UWB
In the DS-UWB technique, the energy of the information
signal is spread by utilizing pseudo random codes. When the
very narrow pulse is applied as a chip waveform, the energy
of the signal is spread over ultra wide spectrum. At a receiver,
a selective rake receiver is applied. The rake receiver
improves the performance by utilizing time diversity, i.e. by
capturing the signal energy propagated through a channel by
different paths and combining the signal components. [2]
Maximum ratio combining (MRC) is an optimal combining
technique by weighting received signal components with
estimated received amplitude and then coherently combining
them. The decision variables in MRC is given by [2]
( τ ) ( ) , 0,1,
1 0
*
,MRC
b
=Σ ∫ − =
=
U a r t w t dt i n i
N
n
T
i n
(1)
where n is the multipath component, an is the complex gain of
the nth multipath component, Tb is the duration of the data bit,
r(t) is the received signal, t is time, τn is the delay of the nth
multipath component and wi(t) is the pulse waveform of the
data bit i.
Equal gain combining (EGC) is similar to MRC, but it does
not weight the signal components with the estimated received
amplitude [2]. Square-law combining (SLC) is non-coherent
approach, where the phase of the signal is considered to be
unknown. This leads to simpler implementation of the
receiver with the cost of the performance. The decision
variables in SLC are [2]
( τ ) ( ) , 0,1.
1
2
0
,SLC
b
=Σ ∫ − =
=
U r t w t dt i
N
n
T
i n i
(2)
The 11th International Symposium on Wireless Personal Multimedia Communications (WPMC 2008)
B. UWB-FM
UWB-FM is a simple technique for UWB communication
with low data rates. It applies a double frequency modulation
(FM): digital frequency shift keying (FSK) with a low
modulation index followed by analog FM with a high
modulation index. This yields to the FM signal V(t) [3]:
( ) sin(ω βcos(ω ) φ ), c m 0 V t = A t − t + (3)
where A is the amplitude, ωc is the angle frequency of the
carrier signal, β is the modulation index, ωm is the angle
frequency of the modulating signal and φ0 is the arbitrary but
time-independent constant.
At the receiver, the FM signal is demodulated with a delayline
FM demodulator. This approach does not need any
frequency translation. Therefore, no local oscillator and no
carrier synchronization are required. This reduces the
complexity of the transceiver. In the delay-line FM
demodulator, the received signal is delayed with time that is
equal to an odd multiple of a quarter period for the carrier
frequency of the FM signal. FSK demodulation can be done
with a bandpass filter followed by a phase-locked loop. [3]
C. Hospital Channel Models
The channel measurement campaign was carried out at the
Oulu University hospital in 2005 [4]. The measurements were
done in three different environments; an operating room
(OR), an x-ray examination room (XR) and an intensive care
unit (ICU) in the frequency range from 3.1 GHz to 6.0 GHz.
The environments were static during the measurements, i.e.
there were no movement inside a room when recording the
data, expect in the ICU. The multipath model was obtained by
investigating the multipath propagated signals in the power
delay profile (PDP). The model was shown to be similar as a
modified IEEE 802.15.3a channel model defined in [6]. The
modified IEEE 802.15.3a parameters for the hospital
environments are presented in Table 1. The example channel
realizations for the environments are illustrated in Figure 1.
The parameters occurred in Table 1 are [6]
• Λ = Cluster arrival rate,
• λ = Ray arrival rate,
• Γ = Cluster decay factor,
• γ = Ray decay factor,
• σ1 = Standard deviation of cluster lognormal fading term,
• σ2 = Standard deviation of ray lognormal fading term and
• σx = Standard deviation of lognormal shadowing term for
total multipath realization.
Table 1: Modified IEEE 802.15.3a model parameters for
hospital.
Model
parameters
Operating
room (OR)
X-ray
room (XR)
Intensive care
unit (ICU)
Λ [1/ns] 0.04 0.05 0.09
λ [1/ns] 2 1.5 2
Γ [ns] 9 13.3 16
γr [ns] 8 10 5
σ1, σ2 [dB] 3.4 3.4 3.4
σx [dB] 1.5 1.5 1.5
Figure 1: One channel realization for each environment.
D. Applications
Electrocardiography (ECG) is an essential tool for
investigating cardiac arrythmias. It is also useful in
diagnosing cardiac disorders such as myocardial infarction
[7]. The standard ECG measurement contains information
from 12 leads. The six chest leads (V1 to V6) measure the
activity of the heart in the horizontal plane. The activity in the
vertical plane is measured by the six limb leads (I, II, III,
aVR, aVL and aVF). When the sample rate of 1250 samples
per second and the resolution of 12 bits/sample are applied,
the total data rate (Rb) from the ECG measurement is 180
kbps [8].
The physiological properties of the muscles at rest and in
contraction are evaluated and recorded by using a technique
called electromyography (EMG). In a surface EMG,
measurements are done non-invasive and it can be applied to,
e.g., childbirth to measure the contraction intervals. In the
surface EMG, the signals from reference and two detection
electrodes construct the EMG signal [9]. Each of the electrode
produces the information rate of 600 kbps with the sample
rate of 50 samples per second and the resolution of 12 bits per
sample [8]. Hence, the total data rate is 1.8 Mbps.
A possible application for x-ray examination is to transmit
wirelessly the taken images from a measurement device to a
viewer, e.g., to computer. Since uncompressed x-ray images
are very large, the wireless link should be fast enough. The
The 11th International Symposium on Wireless Personal Multimedia Communications (WPMC 2008)
data rate for the x-ray imaging application is chosen to be 24
Mbps.
III. SIMULATION CONFIGURATIONS
In order to evaluate the suitability of the DS-UWB and UWBFM
systems for the link between the patient monitoring
sensors and the monitoring devices, software simulators were
implemented in MatlabÓ. In this section, the simulation
parameters and configurations are presented and justified. The
parameters for the systems are chosen so that the systems are
occupying approximately the same frequency band. The
spectrum allocations of the systems with defined parameters
are depicted in Figure 2.
Figure 2: Spectrum allocations of UWB-FM and DS-UWB
with defined parameters.
The chip waveform for DS-UWB was chosen to be 10th
derivative of the Gaussian monocycle with the pulse length
(Tp) of 0.7 ns, thus having the center frequency (fc) of 4.57
GHz and bandwidth (W) of 3.11 GHz. The studied receiver
algorithms for DS-UWB are chosen to be coherent MRC and
EGC and non-coherent SLC with 8-finger selective rake
receiver. The selective rake receiver with 8 fingers has been
shown to be adequate in order to trade-off between
complexity and performance [10]. The modulation scheme for
coherent detection algorithms is binary pulse amplitude
modulation (BPAM) and on-off keying (OOK) for noncoherent
detection algorithm, respectively. BPAM cannot be
applied for SLC, because the polarity of the data bit is needed
in the decision. In OOK, nothing is transmitted in the case of
bit "0". The pulse power of OOK is twofold compared to
BPAM to attain same average transmitted power. According
to the data rates, presented in Section II, pulse repetition
processing gains are adjusted. The pulse repetition processing
gain in decibels is defined as
Therefore, the processing gains for 1.8 Mbps and 24 Mbps are
29.0 dB and 17.7 dB, respectively. With the data rate of 180
kbps and the pulse length of 0.7 ns, the processing gain is as
much as 39 dB. This requires a huge calculation capacity, and
therefore it is not feasible for the medical sensor applications.
In the UWB-FM technique, the channel bandwidth is
adjusted with the modulation index of FM. The UWB-FM
system is studied with the channel bandwidths of 500, 1000,
2000 and 2900 MHz. The centre frequency of the carrier is
4.55 GHz. Through the simulations, the modulation index of
FSK is set to one.