For each frame, tracks are formed for
possible objects and their log likelihood value is calculated.
Each new track that is formed becomes
a new hypothesis.
The log likelihood value is the ratio
of the probability that the track hypothesis belongs to an actual target,
to the probability that it is composed of false alarms.
The deadline for the frame is the arrival
of the next frame.