GMPE340 Kalman Filter and Sensor Fusion

Autonomous vehicles provides the technology towards a more sustainable transportation. The performance of modern transportation systems has been greatly improved by the rapid development of connected and autonomous vehicles, of which high precision and reliable positioning is a key technology. This is one application that relies on the use of sensor fusion with the Kalman filter as the main workhorse providing reliable estimates of the system states (e.g. position, velocity and orientation).

This course introduces the use of stochastic processes and applied Kalman filtering with focus on positioning, navigation and timing applications (PNT).

The first part includes the essential notions of probability, an introduction to random signals and response to linear systems, state-space modelling and Monte Carlo simulations.

The second part contains the main theme of the course, which is applied Kalman filtering. This part starts with the basic filter derivation using the minimum mean square error approach. This is followed by various approaches to the base theory such as: the information filter, suboptimal analysis, conditional density viewpoint, Bayesian estimation, relationship to least squares adjustment (LSQ) and other estimators, smoothing and methods to deal with non-linearities.