Tracking

The useful question in multi-object tracking is not whether the tracker draws a smooth path. It is whether the estimator stays consistent as measurements become partial, noisy, or uneven. This article reads one default run through the diagnostics exported by the filter.

A plain-language argument for multi-source vessel tracking, connected to the design choices behind the sensor-fusion project.

Sequential Monte Carlo, importance sampling, degeneracy, and resampling.

Archived note defining tracking, state variables, and the difference between observations and raw sensor data.

Note on Kalman filtering assumptions, EKF linearization, and simulated tracking performance.

Notes on radar, GPS, AIS, and camera observation models for maritime tracking.

Archived note on the single-target and multi-target simulation setup used to test EKF, PF, and PDA.

Note on Constant Turn Rate and Velocity (CTRV) dynamics and the discrete vessel motion model used for tracking.

Planar vessel state vectors in Cartesian and polar form.

Notes on geodetic, ENU, ECEF, body, seakeeping, and tracking coordinate systems.