Skip to main content

> ML_ALGORITHM // KALMAN-FILTER-EXTENDED-UNSCENTED_v1.0

Kalman Filter (Linear, EKF & Unscented UKF)

The optimal recursive Bayesian state estimation algorithm for linear dynamical systems subject to Gaussian noise, widely used in guidance and navigation.

State Space Dynamic Estimationbayesian-probabilistichigh-intrinsicsmall (<1k)
Back to All Algorithms
Computational Complexity
Training Complexity:Recursive online updates
Inference Complexity:O(state_dim^3) matrix inversion per timestep
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:low
Interpretability & Data
Interpretability Tier:high-intrinsic
Training Data Needs:small (<1k)

Interpretability Assessment

Kalman Gain explicitly quantifies how measurement surprise balances against prior model uncertainty.

Suitable Tasks & Supported Modalities

Suitable Tasks:
state estimationsensor fusiontarget tracking
Supported Modalities:
time-series

Implementing Libraries

filterpy
SciPySciPy Community / NumFOCUS · v1.14.1
View Spec

Foundational Literature

Common Pitfalls & Warnings
  • Extended Kalman Filter (EKF) first-order Taylor expansion diverges when transition non-linearities are severe (use UKF instead)