Theoretical Background

The course assumes basic knowledge of probability theory and Bayesian estimation theory. The course Sensor Fusion (TSRT14) provides a solid background for this course, and is highly recommended to get the most out of this course. Some references to the course material in TSRT14 are provided below to help you pick up the most vital concepts, and you should be able to pick up the basics along the way.

You are assumed to know about the following concepts:

• Bayeisan filtering and the Kalman filters and variantions:
  • Bayes Filtering Recursion (10:18)
  • Kalman Filter (15:39)
  • Kalman Filter Properties (9:25)
  • Extended Kalman Filter (EKF) (12:47)
  • Unscented Kalman Filter (UKF) (11:58)
  • Application: Kalman filters (18:09)
• Common motion models:
  • Continuous Time Motion Models (8:09)
  • Discretizing Motion Models (11:50)

The following concepts are also helpful:

• Detection theory (11:21)
• Filtering CRLB (9:59)
• Filter Banks (21:10)

Coding background

The computer exercises will require some coding skills. Template code and data will be supplied in MATLAB format, but you may use any language of choice to solve the problems. If you use MATLAB, you might find the Sensor Fusion and Tracking toolbox (manual) useful. The toolbox design could also serve as a inspiration for your own code.