|
Toyon has developed a unique approach to feature-aided tracking
in order to manage the effects of kinematic uncertainty in multi-target environments. During kinematically
ambiguous periods of surveillance, track switches and track breaks become prevalent as the tracker
attempts to sort out the association of multiple sensor
measurements to multiple tracks that may be tightly clustered in the surveillance region. Mistakes
made by the tracker during these periods of ambiguity can disastrous to the overall tracking
system as collected signature databases become corrupt and the identity of targets-of-interest (TOIs)
becomes uncertain. Toyon's Tracked Object Manager (TOM) algorithm supervises
the association decisions of the tracker to detect and correct mistakes as they occur. To achieve this,
the TOM exploits signature data collected "on-the-fly" to fingerprint targets during periods of unambiguous
surveillance. Following periods of ambiguity, this signature data is used to resolve ambiguities and
to correct the potential mistakes made by the tracker.
Toyon has applied the TOM to multiple challenging tracking problems and has
demonstrated vast improvements in the ability to maintain pure and continuous tracks as well as pure target signature databases
during periods of tight target spacing and significant gaps in sensor coverage.
For more information regarding Toyon's Tracked Object Manager,
please send queries to ksullivan@toyon.com.
|
