| Summary: | Nowadays, personal positioning systems are more
necessary to build many location-based services. Pedestrian Dead
Reckoning (PDR), which is a pedestrian positioning technique
using the accelerometer sensor to recognize pattern of steps, is an
alternative method that has advantages in terms of
infrastructure-independent. However, the variation of walking
pattern on each individual will make some difficulties for the
system to detect displacement. This is really interested authors to
develop a sensor-based positioning system that applied generally
to all individuals. In the test, 15 test subjects was taken with the
distance of each 10m, 20m and 30m.
Experiment begins with the feasibility test of accelerometer
sensor. In this work, a smartphone with average sampling rate
63.79 Hz and standard deviation of 1.293 is used to records the
acceleration. Then, the acceleration data are analyzed to detect
step and to estimate the travelled distance using several methods.
Detection of steps are able to make an average error of 2.925%,
while the most nearly correct displacement estimation is using
Scarlet experimental method which is make a distance average
error of 1.39metres at all the traveled distance.
|
|---|