Database Reference
In-Depth Information
to factors that affect signal in a nonlinear way, for example the presence of
objects along the signal path, whereas omnidirectional signal attenuations are
not relevant.
Distance and Direction-Based Methods
Unlike the methods described in the previous paragraph, distance-based meth-
ods use absolute RSSI to directly compute distance estimations. This method,
however, requires calibration of the specific radio used, and it is particularly
sensitive to any kind of perturbation. For this reason it is best suited for short-
range distance measurement with known devices. A different approach, based
on the use of a particular array of antennas that are able to compute RSSI for
different directions, uses both the angle of arrival (AOA) and the RSSI to detect
the position of the device. In both cases, assuming no obstacle, the uncertainty
derives from the original observational errors involved in the computation of the
indirect position measure, namely AOA and RSSI errors.
5.3 Uncertainty Models for Spatio-Temporal Data
We now turn to the problem of studying uncertainty of the trajectory of a
moving object. A moving object's trajectory is obtained from raw trajectories,
which are finite sequences of time-space points. The most-used technique for
reconstructing trajectories from trajectory samples (see Chapter 2 in this topic)
is linear interpolation. However, it relies on the assumption that in between
sample points, an object moves at constant minimal speed. It would be more
realistic to assume that moving objects have some physically determined speed
bounds. Given such upper bounds, uncertainty models have been proposed to
estimate the possible positions between every two consecutive points in a trajec-
tory sample. Note that uncertainty in trajectory databases may also arise from
other sources, discussed earlier in this chapter, and also in Chapter 2 .Inthis
way, uncertainty not only refers to the possible locations of a moving object
between two points in a sample, but also to these points themselves, which are
not (in general) exactly recorded. Although the trajectory of a moving object
has been traditionally modeled as a polyline in a 3D space (two dimensions
for geography and one for time), modern approaches model such trajectory as
a volume in 3D, either cylindrical or of a more complex kind. More precisely,
in some of these approaches, the uncertainty of the moving object's position
in between sample points is studied using space-time prisms . Informally, the
space-time prism between two consecutive sample points is defined as the col-
lection of space-time points where the moving objects may have passed, given a
speed limitation. Geometrically, it is the intersection of two cones in the space-
time space such that all possible trajectories of the moving object between the
Search WWH ::




Custom Search