State Estimation
Filters, observers, calibration, mapping, tracking, and sensor fusion.
Filtering Methods
Kalman filter
Estimates linear Gaussian systems optimally in the least-squares sense.
Extended Kalman filter
Linearizes nonlinear dynamics and measurement models locally.
Sigma-point filters
The unscented Kalman filter propagates selected sample points through nonlinear models.
Particle filters
Approximate arbitrary state distributions with weighted samples.
Information filters
Represent uncertainty with information matrices, useful in sparse or distributed estimation.
Square-root filters
Propagate covariance factors to improve numerical conditioning.
U-D filters
Use unit upper-triangular and diagonal covariance factors for stable filtering.
H-infinity filters
Estimate states under worst-case disturbance models instead of stochastic assumptions alone.
Kalman-Bucy filter
Continuous-time Kalman filtering for linear systems driven by stochastic models.
Constrained filters
Enforce known bounds or equality constraints on state estimates.
Smoothers
Fixed-lag or RTS smoothers estimate past states using later measurements.
Wiener filtering
Estimates signals from noisy measurements using second-order statistics.
Recursive least squares
Estimates fixed or slowly varying parameters from streaming data.
Covariance tuning
Consistency checks keep filter uncertainty aligned with observed residuals.