Nonlinear State Estimation: Extended vs Unscented Kalman Filters, Particle Filters, and Factor-Graph Smoothing (iSAM2)

Deep End Era. Engineering with Sankalp · Oct 29, 2026 · 1h 48m

A range-and-bearing measurement can produce a curved posterior that a Gaussian approximation struggles to represent. This episode compares EKF and UKF methods, particle filtering and full-trajectory smoothing, then develops factor graphs, sparse elimination, the Bayes tree and iSAM2.

Chapters

00:00:00 Chapter 1: Cold Open — The "Banana" Distribution of Range-Bearing Sensing and the Four Eras of Nonlinear State Estimation!

00:06:14 Chapter 2: The Bayesian Filtering Recursion — Chapman-Kolmogorov Prediction, Bayes' Rule Update, and Why Nonlinearity Breaks Gaussian Closure!

00:16:58 Chapter 3…