Information Technology Reference
In-Depth Information
SNR
d
the distances between the different points on the
body during movement. A simplified diagram-
matic representation of the proposed system's
data acquisition approach is shown in Figure 1.
The data acquisition procedure is divided into
two phases, namely the initial phase and core
phase. The initial phase is carried out through the
measurement of the subject's height, weight, etc.,
as shown on the right in Figure 1. The aim of the
core phase is to acquire the ranging data between
the different nodes while the subject is walking
through the estimation of the time-of-arrival
(TOA) of the first path, which is then converted
to a distance estimate, as illustrated by the subject
to the left in Figure 1. The system is designed
based on a target ranging accuracy of ≈ 0.1 cm.
This value was particularly chosen for achieving
a ranging accuracy that is ten-times better than
the inter-marker distance accuracy reported in the
literature for current systems. Specifically, the
reported accuracy for the inter-marker distance
for current systems is equal to 1.17 cm (Menz,
Latt et al. 2004; Barker, Freedman et al. 2006).
Generally, applications that are based on wear-
able devices have the low-power consumption
constraint, as they are placed close to the subject's
tissue, to avoid overheating or burning hazard for
the human tissue near transmission (Jovanov 2005;
Bindra 2008). Moreover, accurate gait analysis,
in particular, requires a high ranging accuracy. In
order to choose suitable receiver architecture for
this application, it should be capable of satisfying
the high ranging accuracy and low-power con-
sumption requirements. When transceivers are
compared based on the ranging accuracies that
they can provide, at high SNR, ED estimators
exhibit a floor of
T
s
/ 12
, where
T
s
is the inte-
gration window. Stored-reference estimators,
based on matched-filtering (MF), have perfor-
mance which approaches the CRLB with a be-
havior depending on the fading of the first path.
Typically, EDs require minimal integration win-
dows equivalent to integer multiples of the pulse
2
P z
( )
≅
Q
( )
z
(13)
min
min
2
where
N
ch
is the number of channel realizations,
SNR
is the total signal-to-noise-ratio,
d
,
(*)
is the minimum normalized
distance between the symbols used for the repre-
sentation of the PPM scheme,
i
( ) =
z
d
( )
z
min
min
k i
k
arg
min
,
and
k
is the argument of the minimization (Darda-
ri, Chong et al. 2006).
It is worth noting that this bound provides a
lower-bound on the TOA of coherent detectors.
Typically, the MF estimator performance tends to
the CRLB with a behavior depending on the fading
of the first path (Dardari, Chia-Chin et al. 2008).
Most of the available motion capture and
movement tracking systems are based on the ac-
quirement of the absolute positions of the different
nodes. These positions are then used for the esti-
mation of the other gait parameters. The relevant
approach using UWB radios is the node positioning
approach. Typically, UWB positioning-accuracy
is based on the ranging accuracy. The CRLB on
the positioning accuracy is (Gezici, Tian et al.
2005; Patwari, Ash et al. 2005):
(*)
=
d
2
( )
z
k i
,
i
c
( )
p
≥
(14)
var
2 2
π
SNR
β
w
where, var(.)is the variance,
p
is the estimated
node position,
c
is the speed-of-light, and
β
w
ef-
fective (or root mean square) signal bandwidth.
OVERVIEW OF PROPOSED SYSTEM
In this section, the proposed system is introduced,
and numerical results are given. The proposed
system is based on wearable UWB radios at-
tached to the subject's body, or possibly sewn into
clothing specifically designed for this application.
The target of the proposed system is to acquire
Search WWH ::
Custom Search