Fusion in Signal Processing - Information Fusion in Signal and Image Processing - page 18

windowing for plot track
associations

3
synchronous
observations

center

Figure 2.2. Multi-sensor tracking of a maneuvering target

model, followed by an update phase, based on observed data, that can be implemented
using an extended Kalman filter.

If the different models used are close to reality and if the partial decisions (data
validation choosing an evolutionary model, associating validated data with a track)
were right, then the problem consists of combining several distributions involving a
quantity in order to infer a plausible value for the resulting distribution which takes
into account all of the imprecisions. If one of the partial decisions is incorrect, a con-
junctive or weighted mean combination no longer has any physical meaning and more
elaborate mechanisms are required to account for the problem's uncertainties. This
last comment shows the importance of windowing mechanisms, which are designed
to prohibit combination when the distributions are incompatible. This is relevant to
fusion techniques only insofar as we are discussing the robustness of a mechanism
with respect to modeling defects, the use of not perfectly reliable data, to taking into
account uncertainties which are native or induced by a series of partial decisions.

The distribution combination mechanism may or may not authorize the combina-
tion of distributions, depending on the scenario.

A first windowing mechanism authorizes the combination:
- when one of the distributions is much more precise than the other, the combina-
tion must then behave as a conjunctive mechanism, such that the resulting distribution
generally behaves like the most precise of the distributions;