General EKF implementation (non-SLAM)

Search IconIcon to open search

Parent: Extended Kalman Filter

Source: SLAM for Dummies General (non-SLAM) implementation of EKF:

  • only state estimation
  • robot is given a perfect map
  • no map update necessary

SLAM implementations of EKF requires map update and therefore the matrices are changed.

Source: Scaradozzi 2018 SLAM application in surgery EKF vs KF

  • circumvents linearity assumption
  • uses nonlinear functions to describe
    • the next state probability
    • measurement probability
  • approximates the state distribution with a Gaussian Random Variable