Observer theory and non-linear design techniques, 6 credits
Graduate course at Vehicular Systems. The nominal number of credits for the course is 6. Course information, courlse literature, and latest lecture schedule can be found at the course git-repo https://gitlab.liu.se/vehsys/nonlinearobservers_phd
Course objectives, organization, and exaination
The objective of the course is summarized as:
- give fundamental knowledge in basic notions for observers. This is done for linear and non-linear models in both ODE and DAE formulations.
- give basic knowledge on some of the most common design methods used for observers, primarily for non-linear systems.
- get some experience in implementing different types of observers and apply to real and simulation cases.
The course is organized in lectures and discussion seminars. Participation is mandatory for passing the course.
Course syllabus
- Basic notions, linear systems
- observability for ODE and DAE
- partial observability
- observability index, observability indices
- how to test for observability
- model formulations particularly well suited for observer design
- geometric interpretations
- observability gramians
- Nonlinear extensions of linear notions
- Methodology - common knowledge
- Stochastic filters
- Kalman Filter
- Extended Kalman Filter
- Unscented Kalman Filter
- Particle Filter
- High gain observers
- Sliding mode observers
- Direct Lyapunov design
- Constant-Gain EKF, Gain-scheduled observers
- Stochastic filters
- Implementation of observers
Informationsansvarig: Per Öberg
Senast uppdaterad: 2021-02-22