A Full-Body Wireless Wearable UWB-Based Human Motion Capture and Gait Analysis System - E-Healthcare Systems and Wireless Communications - page 450

Figure 4. Proposed node grouping into LOS regions assuming the Vicon marker-set (Shaban 2010;
Shaban, El-Nasr et al. 2010) ©2010 IEEE. Reprinted, with permission, from Shaban et al., “Toward a
Highly Accurate Ambulatory System for Clinical Gait Analysis via UWB Radios”, in IEEE publication
title: IEEE Transactions on information technology in Biomedicine, vol. 14, NO. 2, Mar. 2010.

signals to each of the four reference off-body nodes
to measure the TOA at a pre-assigned time-slot.
Typically, the on-body nodes are synchronized
through wire connections. We assume a minimal
number of wires to guarantee freedom of move-
ment. Similarly, reference nodes are synchronized
through wire connections. In order to synchronize
on-body nodes and reference-nodes, we assume
a wire connection between the pad, where the
reference nodes are fixed, and the on-body nodes
through a plug. This plug is connected during the
system initialization, and is removed afterwards.
We assume that the system initialization is
subdivided into initial and subsequent frames. The
initial frame requires 0.14 seconds and 2e4 subse-
quent frames each of which is 1 ms. This leads to
an overall system initialization of 20.14 seconds.
In the initial-frame of the system initialization
procedure, we assume an active time per node of

2.8 ms, which is subdivided into ten time-slots.
At each time slot, twenty pulses are averaged in
order to obtain a processing gain, as well as higher
ranging and positioning accuracies. Then, these ten
time slots are averaged in order to obtain a finer
range estimate. This initial-frame is followed by
2e4 subsequent-frames. After the acquisition of the
TOA of the initial-frame, subsequent-frames do a
fine tracking of acquired TOA values. Particularly,
at each subsequent-frame we search over five
samples around the previous TOA to determine
the current peak. This provides us with the TOA
offset w.r.t. the preceding value. We also assume
averaging of twenty pulses per TOA estimate for
subsequent-frames. These obtained ranges are
further employed in a linear least-squares (LS) lo-
calization approach, which is typically performed
at an off-body centralized node.