Course content

The course is taught evry year, next time in the Autumn 2010. Bayes estimation in nonlinear, non-gaussian systems:
Nonlinear discret-time state space models (unscented Kalman filter, point-mass and particle filters). Hidden markov models (HMM) (forward, backward, Viterbi and Baum-Welch algorthms). Static and dynamic multiple models (MMEA, GPB1, GPB2 and IMM algorthms).
The methods will applied to navigation and tracking problems.

Learning outcome

Learn the latest methods used for estimating a vehicles position, velocity and attitude in a modern navigation system.

Learning methods and activities

Lectures, problem sets and project assignments.

Compulsory assignments

• One term project

Recommended previous knowledge

TTK4180 or TTK4605

Required previous knowledge

Design of robust Kalman filters, parameter estimation in linear state space models, applying the Kalman filter to nonlinear systems (linearized and extended Kalman filters).

Course materials

Notes and textbook