Exact probabilistic tracking for radar systems.
Riverside Research
Resolve is measured head to head against the Kalman filter, the standard for sixty years, in post-mission analysis on the same data. It produces the better track.
Both start from the same returns. As the noise rises, the Kalman filter's track error climbs with it. Resolve's stays low, because it weighs every return it has seen and every one that follows against the physics of flight and the geometry of the radar, and reports where the target is with an exact probability attached. A sensor upgrade delivered in software.
Tested with Riverside Research. The curves are illustrative. Post-mission analysis today. Real time is on the roadmap.
Resolve starts from one probabilistic model of the target's motion and the radar's geometry. Search decides which return belongs to which object. Infer reasons over the whole track and returns an exact probability for every position.
Decides which return belongs to which object across time, so a sharp turn does not start a new track.
Reasons over the whole track, backward and forward, and returns an exact probability for every position.
Feed it a recorded run. It computes the track, with an exact probability on every position. Real time is on the roadmap.
Recorded returns in, one exact track out, with a probability on every position.
The same model, running as the returns arrive.