at which there is a real change of direction (critical points), mainly
due to errors in GPS data. Erroneous identification of critical points
is fatal for navigation of visually impaired.
Another major problem in obtaining a trace of route is a large
amount of memory needed to record the track, especially when
walking. This imposes a restriction on the number of saved tracks
and the maximum number of points by which they are described.
For example, Garmin GPS navigators for pedestrians, which have
not supported GPS maps, can save 10 or 20 tracks, and each track
can contain up to 10,000 points. The solution for this problem is
to use algorithms for real-time track simplification.
Since each track is described by a sequence of points, it can be
assumed that track simplification algorithms are part of the line
simplification algorithms. McMaster (1987) classifies them into
five categories: (1) Independent point algorithms, (2) local processing algorithms, (3) Unconstrained extended local processing
algorithms, (4) constrained extended local processing algorithms
and (5) global processing algorithms. The first four types of
algorithms belong to Local line simplification algorithms, and
the last —to the Global line simplification algorithms (Shi and
Cheung, 2006).
Local simplification algorithms: The relationship between every
two or three consecutive points is analyzed. These types of
algorithms are very simple and fast, but because of the local data
processing it is very difficult to obtain optimal results. More
commonly used local track simplification algorithms are as
follows: nth point – only one point between n consecutive points
is retained; distance threshold based algorithms – all points, for
which the distance to the preceding track point is less than the