Localization with Extend Kalman Filter

DESCRIPTION: Homework 5 of Mobile Robotic
AUTHOR         : Tsung-Han Brian Lee
LICENSE         : MIT

---

Handling Multiple Observations

Parallel EKF

Updating the mean and covaraince jointly by augumenting the measurement vector and augumenting the Jacobian of sensor model.

e.g. 1 observation measurement shape = (2, 1), while 3 observations measurement shape = (6, 1).

Sequential EKF

Updating the mean and covariance of state separately and sequentially.

---

Get Start

Simulate under unkown correspondency case

16 KNOWN_CORRESPONDENCY   = false;

Change Watch Scope

17 WATCH_SCOPE            = <value_you_want>;

Plot Covariance with different confident

In plotGaussian.m, modify to meet your need.

33 chisquare_val = sqrt(chi2inv(<confidence>, 2));

---

REFERENCE

• Dynamic Systems amd Control Lab @ National Tsing Hua University